Should the AI ask before it acts?
Auto-act on high-confidence, reversible steps; ask first on uncertain or costly ones; always offer a quick dismiss.
What the research says
Acting automatically only pays off when the system is confident and the action is cheap to undo. When it's unsure, asking (or doing nothing) beats guessing, and users should be able to invoke or dismiss the AI easily.
“If a system is uncertain about a user’s intentions, it should be able to engage in an efficient dialog with the user, considering the costs of potentially bothering a user needlessly.”
Horvitz, 1999
Sources
- Principles of mixed-initiative user interfaces1999 · CHI 1999 · Eric Horvitz
“If a system is uncertain about a user’s intentions, it should be able to engage in an efficient dialog with the user, considering the costs of potentially bothering a user needlessly.”
Principle 5 - Guidelines for Human-AI Interaction2019 · CHI 2019 · Saleema Amershi, Dan Weld, Mihaela Vorvoreanu, et al.
“Make clear how well the system can do what it can do. Help the user understand how often the AI system may make mistakes.”
Guideline G2